Molecular networks are the information processing devices of cells and organisms, transforming signals into coherent cellular responses. Networks are remarkably flexible and can re-configure in an adaptive response to perturbation. This ability is apparent at all levels - from fast epigenetic changes in response to environmental signals or developmental cues, to re-organization to accommodate pathological changes in cancer, to genetic changes underlying network evolution under selection. Reconfiguration is essential to networks' function as well as to their ability to evolve new functionality. We understand little, however, about how specific genetic and epigenetic changes allow novel functions to emerge in complex networks. Genomics has recently made it possible to collect massive datasets about temporally changing systems. Among molecular systems, regulatory networks controlling gene transcription are the most accessible for systems-scale analysis. The availability of scalable, cost-effective genomics approaches to systematically perturb and measure all levels of a transcriptional response along with sophisticated computational methods offer an extraordinary opportunity to study network function. We propose to develop a novel integrated experimental and computational framework to systematically decipher how regulatory networks assume novel adaptive functions through fast epigenetic changes or slow genetic changes. We will distinguish two types of temporal processes. For linear trajectories we will study epigenetic reconfiguration following a nutritional change, and genetic reconfiguration in yeast and cancer cells under selection. For lineages we will characterize the development of novel transcriptional states in the hematopoiesis ontogeny and the evolution of regulatory networks in the Ascomycota phylogeny. The work will unify disparate problems - including how cell adapts to changing growth conditions, how cancer develops, and how species evolve - under a single theoretical and methodological framework. It will help establish a new paradigm for genomics research by moving us from a static snapshot view to a fully dynamic perspective on molecular processes.